718 Introduction

I have now produced a few demos on the 718, data-limited version of POSEIDON. The general summary of the work is that:

  1. If we assume \(\frac M K\) is distributed according to Fishlife priors, the fishery is already quite likely to be sustainable for L. malabaricus (but not A. brevis)
  2. If we assume \(\frac M K\) is low (below 0.9), the fishery is likely to be at a depleted stage already for all species and will likely get worse in the short run as effort from the depleted A. brevis is re-directed at L. malabaricus

Here I focus on the second scenario and in particular I run a management-strategy evaluation using the LBSPR policy suggested in Hordyk, Loneragan & Prince (2015) to change effort.
More precisely, we measure SPR for L. malabaricus and we reduce total allowable effort \(E_t\) whenever SPR is below 40% through the formula: \[ v_t = 0.3 \times \left( \frac{SPR_t}{40\%} - 1 \right)^3 + 0.05 \left( \frac{SPR_t}{40\%} - 1 \right) \] \[ E_{t+1} = E_t v_t \] Effort is never allowed to increase or decrease by more than 10% (i.e. \(v_t \in [0.9,1.1]\)) year on year.

Where \(SPR_{\text{Target}}\) is 40%. We compute SPR by analyzing the catches made by 18 local long-liners, 16 long-range long-liners and 4 gillnetters, as in the CODRS data-set until DEC 2019, and TNC formula (i.e. linear selectivity).

We perform the MSE over 181 separate parameterizations of the model, each representing an “accepted” fishery, roughly congruent with current landings and SPR trends.
For each we checked the long-term (30 years) effect of imposing the LBSPR based policy described above for different assumed values of \(\frac M K\) (which we assume will remain unknown).

Effort in POSEIDON translates to number of days the fishery is open (“365” initially).

MSE Results

Much like the single-species demo, the main take-aways are:

  1. The “ultimate goldilocks” (which I think is a phrase from Tom): set \(\frac M K\) too high (\(\approx 2\)) and you are basically at business as usual, set it too low and you will cut effort forever (ironically making less landings in the long-run than BAU).
  2. Controlling L. malabaricus controls the other species as well, but imperfectly. For L. laticaudis long-term landings go down, this is because it was already doing well so reducing effort will matter little for regrowth and will just mean less yield. For A. brevis the effect is limited because it is expensive enough that people keep fishing it
  3. Most policies involve 10% reduction in effort year after year for at least a decade. This is somewhat dampened by many boats that had previously quit or had become “seasonal”, coming back as stocks rebound (so that effort decreases more slowly than season length). This however just prolong the time the fishery spends at SPR below 40%
## [1] 25
## [1] 40

Does it matter if we use TNC vs Hordryk formula?

In practice, the TNC formula is slightly more pessimistic than the Hordyk one, i.e. it outputs slightly lower SPRs (at all \(\frac M K\) and regardless of whether gillnetters are added in the sample or not).
The main problem is that while the SPR itself is only a few percentage points lower, because it is consistently so over the years you end up with large changes in depletion and effort in the long run.

Overall this means that TNC formula is more risk averse: lower effort higher depletion in the long run.
In terms of landings the effect is mixed: lower landings if the \(\frac M K\) was low but higher landings otherwise (stocks recovered faster).